Skip to content

Language Review

A review procedure — instantiates Social Reality Construction Audit

Walks the live forms, scripts, interfaces, policies, and training materials for the wording and category assumptions that quietly stabilize a constructed reality, and flags what to change.

A Language Review is an operational sweep of the organization's working artifacts — the intake form, the dropdown menu, the call-center script, the onboarding deck, the policy boilerplate. It looks for the assumptions and category effects baked into the language people actually transact through, and it is oriented to producing a flag list for revision. Unlike discourse analysis, which interprets a corpus of rhetoric to theorize how meaning is made, a language review is a practical inspection of live touchpoints, tying each flagged phrase to the artifact and system that keeps propagating it.

Example

A reviewer walks a national census-and-immigration form field by field. A "race/ethnicity" question offers fixed checkboxes and no write-in. A "head of household" field presumes there is exactly one. A "marital status" list is missing options many respondents need, and the enumerator instructions say to "pick the closest." Each of these choices quietly enforces a category rather than recording one. The review flags where the form manufactures the very categories it claims to observe, links each flag to the field and system that carries it, and marks candidates for revision. A form field, in Austin's terms, is a performative: checking a box does not describe a pre-existing fact, it enacts a classification.[1]

How it works

The distinguishing discipline is to inventory live artifacts, read them for embedded assumptions, tie each to its carrier, and flag:

  • Collect the operational language — forms, scripts, interfaces, policies, training materials.
  • Read for stabilizing assumptions — forced categories, loaded defaults, required fields, presupposing wording.
  • Map each flag to its carrier — the database field, the policy clause, the UI element that reproduces it.
  • Flag revision candidates — and hand them off; the review recommends, it does not decide or execute.

Tuning parameters

  • Artifact scope — public-facing copy only versus the full operational stack, including internal scripts and database codes. Deeper scope finds the assumptions baked into infrastructure.
  • Assumption sensitivity — flag only explicit bias versus subtle presuppositions like default values and required fields. A lower threshold catches the quiet ones.
  • Traceability — merely note the wording versus link each flag to the system that reproduces it. Linking is what makes the fix actionable.
  • Flag-versus-fix boundary — how far to go toward proposing replacements, given that the review recommends and the revision executes.

When it helps, and when it misleads

Its strength is catching construction where it actually operates — the required field, the missing option, the forced binary — rather than in speeches, and producing a concrete, fixable list. Its classic misuse is the cosmetic word-swap that leaves the underlying category and its system untouched (rename the label, keep the forced binary), and a related failure is running the review as a compliance checklist with no authority to change anything. The discipline that guards against this is to trace every flag to its carrier system, so a fix has to reach the infrastructure and not just the display text.

How it implements the components

Language Review realizes the operational-artifact side of the archetype — the language embedded in live systems, and the systems that carry it:

  • language_pattern_inventory — the flagged wording, labels, and forced categories found in live artifacts.
  • institutional_reinforcement_map — ties each language flag to the form, field, or policy that propagates it, showing where the wording is load-bearing.

It does not interpret how a discourse produces meaning across a corpus (contested_interpretation_record → Discourse Analysis) or decide and execute the change (revision_or_preservation_decision → Revision Decision Memo; implementation_revision_pathway → Taxonomy or Form Revision).

Notes

It feeds Taxonomy or Form Revision: the review flags the wording and the carrier; the revision changes it. Distinct from Discourse Analysis — live artifacts and concrete fixes here, corpus and theory there.

References

[1] J. L. Austin, How to Do Things with Words (1962) — performative utterances do things rather than merely describe them ("I now pronounce you…"). Named because a form field enacts the classification it appears to record. ↩